| dc.description.abstract | There is a lack of empirical evidence that the dorsolateral prefrontal cortex (DLPFC) is exerting control over post-error selective attention adjustments. This is an error compensation mechanism that is hypothesized by the conflict monitoring theory of error processing. Understanding the error detection and compensation system enables the easily detectable error-related negativity (ERN) EEG signal, which reliably occurs approximately 100 ms after an erroneous action/response, to be utilized in brain-computer interfacing (BCI) systems effectively. An error-aware BCI has many promising applications and advantages, including error mediation in critical tasks such as piloting, error tracking in training programs, as well as allowing more responsive interpretation of commands for disabled users.
This thesis focused on utilizing a speeded-response task in the somatosensory modality to determine whether sensory processing is modulated post error commission, specifically at the stages controlled by the DLPFC. Additionally, we investigated whether the ERN, cognitive control, and temporary behavioural adjustments are linked. We hypothesized that a customized version of the flanker task that applies vibrotactile stimuli could effectively generate the ERN and characteristic post-error behavioural adjustments; and that the early somatosensory event-related potentials (ERPs) will be modulated in post-error trials compared to post-correct trials; additionally, larger ERNs would correspond to a greater degree of modulation and greater reduction of interference in post-error performance. The current results demonstrate that error commission does induce a clear modulation of early somatosensory processing components, which reflect the DLPFC’s influence of spatial selective attention. Furthermore, this effect is conditional on the interpretation of error: ambiguity dampens the sensory cortical modulations.
Subject-wise correlations showed that in the pre-error state, the size of somatosensory ERPs did not predict performance, but post error commission, the temporarily modulated P100 amplitude became correlated with individual performance. Furthermore, post-error P100 amplitudes correlated with participant ERN amplitudes. The degree of change in performance following error commission, however, only correlated (in one aspect) with individual ERN size, and not with the selective attention related signals. These findings suggest that perhaps error detection circuits directly induce some compensatory adjustments onto the current behaviour; while it also recruits the DLPFC, which heightens optimization of selective attention and influences the general strategy to the task.
In general, the combined paradigm of speeded-response flanker task and vibrotactile discrimination task was an effective method to probe the error system. There is potential for more trends to be detected, which could be achieved with a larger participant group, for example, by dividing data into subgroups based on individual compensatory strategy.
In addition to providing support to theories of error processing, a directly utilizable finding of the current study is that people cannot effectively compensate for errors if the circumstances are ambiguous, even if internal error detection was successful. Hence, if a BCI training tool monitors the ERN, it can enhance the user practice experience by explicitly displaying parameters and highlighting errors as they occur. | en |
